Use merge-base diffing in link-check lints

[Copilot]

Summary

link-check / lint-urls was diffing base..head, so PR jobs could lint unrelated URLs introduced on main after the branch diverged. This narrows both URL and xref linting to lines introduced by the PR by diffing from the merge base instead.

• Scope

  • Update scripts/lint_urls.sh to use git diff base...head
  • Update scripts/lint_xrefs.sh to use the same merge-base semantics

• Behavior change

  • Before: lint jobs could inspect files changed only on the base branch
  • After: lint jobs inspect only files/lines reachable from the PR head relative to the merge base

• Regression coverage

  • Add a focused test that creates diverged main and feature histories and verifies only the feature-side changes are linted for:
    • URL checks
    • cross-reference checks

# before
git diff "$BASE..$HEAD"# after
git diff "$BASE...$HEAD"

Test plan

Added regression coverage in .ci/scripts/tests/test_link_check_diff_selection.py for both lint_urls.sh and lint_xrefs.sh.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/21762

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❌ 2 New Failures

As of commit ed0c8cd with merge base e60faa2 ():

NEW FAILURES - The following jobs have failed:

• Build Presets / apple (pybind) / build (gh)
  ##[error]fatal: unable to access 'https://github.com/leetal/ios-cmake/': SSL certificate problem: self signed certificate
• Lint / link-check / lint-urls (gh)

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[github-actions]

This PR needs a release notes: label

If your change should be included in the release notes (i.e. would users of this library care about this change?), please use a label starting with release notes:. This helps us keep track and include your important work in the next release notes.

To add a label, you can comment to pytorchbot, for example
@pytorchbot label "release notes: none"

For more information, see
https://github.com/pytorch/pytorch/wiki/PyTorch-AutoLabel-Bot#why-categorize-for-release-notes-and-how-does-it-work.

[Copilot]

Pull request overview

This PR updates the link-check lint scripts to diff against the merge base (base...head) instead of the base tip (base..head), so URL and xref linting only inspects lines introduced by the PR even when main has advanced since the feature branch diverged.

Changes:

• Switch scripts/lint_urls.sh and scripts/lint_xrefs.sh from git diff base..head to git diff base...head.
• Add a regression test that constructs diverged main and feature histories and asserts only feature-side changes are linted.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

File	Description
scripts/lint_xrefs.sh	Use merge-base diff semantics for PR-only xref linting.
scripts/lint_urls.sh	Use merge-base diff semantics for PR-only URL linting.
.ci/scripts/tests/test_link_check_diff_selection.py	Regression test covering diverged-history selection for both URL and xref linting.

Suppressed comments (2)

scripts/lint_urls.sh:78

• With set -euo pipefail, this per-file pipeline can terminate the whole diff scan when a changed file has no “+” lines (e.g., deletion-only diffs). In that case grep -Ev '^\+\+\+' returns exit code 1, which aborts the process substitution early and can cause the script to miss URLs in later files.

 git diff --unified=0 "$1...$2" -- "$filename" "${excludes[@]}" \
| grep -E '^\+' \
| grep -Ev '^\+\+\+' \
| perl -nle 'print for m#'"$pattern"'#g' \
| sed 's|^|'"$filename"':|'

scripts/lint_xrefs.sh:43

• With set -euo pipefail, this per-file pipeline can terminate the whole diff scan when a changed file has no “+” lines (e.g., deletion-only diffs). In that case grep -Ev '^\+\+\+' returns exit code 1, which aborts the process substitution early and can cause the script to miss xrefs in later files.

 git diff --unified=0 "$1...$2" -- "$filename" "${excludes[@]}" \
| grep -E '^\+' \
| grep -Ev '^\+\+\+' \
| perl -nle 'print for m#'"$pattern"'#g' \
| sed 's|^|'"$filename"':|'

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated no new comments.

Suppressed comments (5)

.github/workflows/_link_check.yml:33

• This workflow still uses the default shallow checkout (actions/checkout without fetch-depth: 0). If lint_urls.sh / lint_xrefs.sh use merge-base diffing (git diff A...B), a shallow clone can be missing the merge base even after git fetch origin "$BASE_REF", causing merge-base-based diffs to fail.

Consider unshallowing the checkout (or setting fetch-depth: 0) so merge-base computations are reliable.

 if [ -n "$BASE_REF" ] && [ "$BASE_REF" != "0000000000000000000000000000000000000000" ]; then
git fetch --no-tags origin "$BASE_REF"
args=("$BASE_REF" "$HEAD_REF")

.ci/scripts/tests/test_link_check_diff_selection.py:104

• Similarly, this xref test won't distinguish BASE..HEAD from BASE...HEAD because main.md is a base-side addition (which becomes a deletion in the diff and is filtered out). Make main remove an xref that exists in the merge base so BASE..HEAD would incorrectly lint it as an added line.

 def test_lint_xrefs_uses_merge_base_for_changed_lines(self) -> None:
base_sha, head_sha = self.create_diverged_history(
"feature.md",
textwrap.dedent(
"""\
[feature](docs/feature-target.md)
"""
),
"main.md",
textwrap.dedent(
"""\
[main](docs/main-target.md)
"""
),
)

.ci/scripts/tests/test_link_check_diff_selection.py:80

• As written, this test won't fail if lint_urls.sh still uses git diff BASE..HEAD: the main.md URL is introduced only on the base side, so it shows up as a deletion (and the script filters to + lines). To make the test distinguish .. vs ..., have main remove a URL that exists in the merge base (so BASE..HEAD would treat it as a new + line in HEAD).

 def test_lint_urls_uses_merge_base_for_changed_lines(self) -> None:
base_sha, head_sha = self.create_diverged_history(
"feature.md",
"https://example.com/feature\n",
"main.md",
"https://example.com/main\n",
)

scripts/lint_xrefs.sh:40

• git diff A...B depends on being able to compute the merge base. In the link-check workflow this script is typically run in an actions/checkout shallow clone (default fetch-depth: 1), and _link_check.yml still fetches BASE_REF shallowly for some jobs, which can leave the merge base unreachable and make this loop fail with "no merge base" errors.

Consider computing the merge base once with git merge-base (so failures can be reported with a clear message) and diffing $merge_base..$2 instead of using ... directly.

 if [ $# -eq 2 ]; then
for filename in $(git diff --name-only --unified=0 "$1...$2"); do
git diff --unified=0 "$1...$2" -- "$filename" "${excludes[@]}" \
| grep -E '^\+' \

.ci/scripts/tests/test_link_check_diff_selection.py:24

• The test setup doesn't currently create any URL/xref in the merge base commit, so it can't detect the difference between git diff BASE..HEAD and git diff BASE...HEAD. To exercise the regression, the merge base should contain a URL/xref that main later removes (so BASE..HEAD would see it as an added line in HEAD, even though it predates the branch).

Without that, these tests can pass even if the scripts still use .. diffing.

This issue also appears in the following locations of the same file:

• line 74
• line 90

 self.write("README.md", "base\n")
self.write("docs/feature-target.md", "feature target\n")
self.write("docs/main-target.md", "main target\n")
self.run_cmd("git add README.md docs/feature-target.md docs/main-target.md")

[psiddh]

Closing this PR in favir of #21765

[Copilot]

Summary

link-check / lint-urls was diffing base..head, so PR jobs could lint unrelated URLs introduced on main after the branch diverged. This narrows both URL and xref linting to lines introduced by the PR by diffing from the merge base instead.

• Scope

  • Update scripts/lint_urls.sh to use git diff base...head
  • Update scripts/lint_xrefs.sh to use the same merge-base semantics

• Behavior change

  • Before: lint jobs could inspect files changed only on the base branch
  • After: lint jobs inspect only files/lines reachable from the PR head relative to the merge base

• Regression coverage

  • Add a focused test that creates diverged main and feature histories and verifies only the feature-side changes are linted for:
    • URL checks
    • cross-reference checks

# before
git diff "$BASE..$HEAD"# after
git diff "$BASE...$HEAD"

Test plan

Added regression coverage in .ci/scripts/tests/test_link_check_diff_selection.py for both lint_urls.sh and lint_xrefs.sh.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/21762

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❌ 2 New Failures

As of commit ed0c8cd with merge base e60faa2 ():

NEW FAILURES - The following jobs have failed:

• Build Presets / apple (pybind) / build (gh)
  ##[error]fatal: unable to access 'https://github.com/leetal/ios-cmake/': SSL certificate problem: self signed certificate
• Lint / link-check / lint-urls (gh)

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[github-actions]

This PR needs a release notes: label

If your change should be included in the release notes (i.e. would users of this library care about this change?), please use a label starting with release notes:. This helps us keep track and include your important work in the next release notes.

To add a label, you can comment to pytorchbot, for example
@pytorchbot label "release notes: none"

For more information, see
https://github.com/pytorch/pytorch/wiki/PyTorch-AutoLabel-Bot#why-categorize-for-release-notes-and-how-does-it-work.

[Copilot]

Pull request overview

This PR updates the link-check lint scripts to diff against the merge base (base...head) instead of the base tip (base..head), so URL and xref linting only inspects lines introduced by the PR even when main has advanced since the feature branch diverged.

Changes:

• Switch scripts/lint_urls.sh and scripts/lint_xrefs.sh from git diff base..head to git diff base...head.
• Add a regression test that constructs diverged main and feature histories and asserts only feature-side changes are linted.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

File	Description
scripts/lint_xrefs.sh	Use merge-base diff semantics for PR-only xref linting.
scripts/lint_urls.sh	Use merge-base diff semantics for PR-only URL linting.
.ci/scripts/tests/test_link_check_diff_selection.py	Regression test covering diverged-history selection for both URL and xref linting.

Suppressed comments (2)

scripts/lint_urls.sh:78

• With set -euo pipefail, this per-file pipeline can terminate the whole diff scan when a changed file has no “+” lines (e.g., deletion-only diffs). In that case grep -Ev '^\+\+\+' returns exit code 1, which aborts the process substitution early and can cause the script to miss URLs in later files.

 git diff --unified=0 "$1...$2" -- "$filename" "${excludes[@]}" \
| grep -E '^\+' \
| grep -Ev '^\+\+\+' \
| perl -nle 'print for m#'"$pattern"'#g' \
| sed 's|^|'"$filename"':|'

scripts/lint_xrefs.sh:43

• With set -euo pipefail, this per-file pipeline can terminate the whole diff scan when a changed file has no “+” lines (e.g., deletion-only diffs). In that case grep -Ev '^\+\+\+' returns exit code 1, which aborts the process substitution early and can cause the script to miss xrefs in later files.

 git diff --unified=0 "$1...$2" -- "$filename" "${excludes[@]}" \
| grep -E '^\+' \
| grep -Ev '^\+\+\+' \
| perl -nle 'print for m#'"$pattern"'#g' \
| sed 's|^|'"$filename"':|'

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated no new comments.

Suppressed comments (5)

.github/workflows/_link_check.yml:33

• This workflow still uses the default shallow checkout (actions/checkout without fetch-depth: 0). If lint_urls.sh / lint_xrefs.sh use merge-base diffing (git diff A...B), a shallow clone can be missing the merge base even after git fetch origin "$BASE_REF", causing merge-base-based diffs to fail.

Consider unshallowing the checkout (or setting fetch-depth: 0) so merge-base computations are reliable.

 if [ -n "$BASE_REF" ] && [ "$BASE_REF" != "0000000000000000000000000000000000000000" ]; then
git fetch --no-tags origin "$BASE_REF"
args=("$BASE_REF" "$HEAD_REF")

.ci/scripts/tests/test_link_check_diff_selection.py:104

• Similarly, this xref test won't distinguish BASE..HEAD from BASE...HEAD because main.md is a base-side addition (which becomes a deletion in the diff and is filtered out). Make main remove an xref that exists in the merge base so BASE..HEAD would incorrectly lint it as an added line.

 def test_lint_xrefs_uses_merge_base_for_changed_lines(self) -> None:
base_sha, head_sha = self.create_diverged_history(
"feature.md",
textwrap.dedent(
"""\
[feature](docs/feature-target.md)
"""
),
"main.md",
textwrap.dedent(
"""\
[main](docs/main-target.md)
"""
),
)

.ci/scripts/tests/test_link_check_diff_selection.py:80

• As written, this test won't fail if lint_urls.sh still uses git diff BASE..HEAD: the main.md URL is introduced only on the base side, so it shows up as a deletion (and the script filters to + lines). To make the test distinguish .. vs ..., have main remove a URL that exists in the merge base (so BASE..HEAD would treat it as a new + line in HEAD).

 def test_lint_urls_uses_merge_base_for_changed_lines(self) -> None:
base_sha, head_sha = self.create_diverged_history(
"feature.md",
"https://example.com/feature\n",
"main.md",
"https://example.com/main\n",
)

scripts/lint_xrefs.sh:40

• git diff A...B depends on being able to compute the merge base. In the link-check workflow this script is typically run in an actions/checkout shallow clone (default fetch-depth: 1), and _link_check.yml still fetches BASE_REF shallowly for some jobs, which can leave the merge base unreachable and make this loop fail with "no merge base" errors.

Consider computing the merge base once with git merge-base (so failures can be reported with a clear message) and diffing $merge_base..$2 instead of using ... directly.

 if [ $# -eq 2 ]; then
for filename in $(git diff --name-only --unified=0 "$1...$2"); do
git diff --unified=0 "$1...$2" -- "$filename" "${excludes[@]}" \
| grep -E '^\+' \

.ci/scripts/tests/test_link_check_diff_selection.py:24

• The test setup doesn't currently create any URL/xref in the merge base commit, so it can't detect the difference between git diff BASE..HEAD and git diff BASE...HEAD. To exercise the regression, the merge base should contain a URL/xref that main later removes (so BASE..HEAD would see it as an added line in HEAD, even though it predates the branch).

Without that, these tests can pass even if the scripts still use .. diffing.

This issue also appears in the following locations of the same file:

• line 74
• line 90

 self.write("README.md", "base\n")
self.write("docs/feature-target.md", "feature target\n")
self.write("docs/main-target.md", "main target\n")
self.run_cmd("git add README.md docs/feature-target.md docs/main-target.md")

[psiddh]

Closing this PR in favir of #21765

[Copilot]

Summary

link-check / lint-urls was diffing base..head, so PR jobs could lint unrelated URLs introduced on main after the branch diverged. This narrows both URL and xref linting to lines introduced by the PR by diffing from the merge base instead.

• Scope

  • Update scripts/lint_urls.sh to use git diff base...head
  • Update scripts/lint_xrefs.sh to use the same merge-base semantics

• Behavior change

  • Before: lint jobs could inspect files changed only on the base branch
  • After: lint jobs inspect only files/lines reachable from the PR head relative to the merge base

• Regression coverage

  • Add a focused test that creates diverged main and feature histories and verifies only the feature-side changes are linted for:
    • URL checks
    • cross-reference checks

# before
git diff "$BASE..$HEAD"# after
git diff "$BASE...$HEAD"

Test plan

Added regression coverage in .ci/scripts/tests/test_link_check_diff_selection.py for both lint_urls.sh and lint_xrefs.sh.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/21762

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❌ 2 New Failures

As of commit ed0c8cd with merge base e60faa2 ():

NEW FAILURES - The following jobs have failed:

• Build Presets / apple (pybind) / build (gh)
  ##[error]fatal: unable to access 'https://github.com/leetal/ios-cmake/': SSL certificate problem: self signed certificate
• Lint / link-check / lint-urls (gh)

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[github-actions]

This PR needs a release notes: label

If your change should be included in the release notes (i.e. would users of this library care about this change?), please use a label starting with release notes:. This helps us keep track and include your important work in the next release notes.

To add a label, you can comment to pytorchbot, for example
@pytorchbot label "release notes: none"

For more information, see
https://github.com/pytorch/pytorch/wiki/PyTorch-AutoLabel-Bot#why-categorize-for-release-notes-and-how-does-it-work.

[Copilot]

Pull request overview

This PR updates the link-check lint scripts to diff against the merge base (base...head) instead of the base tip (base..head), so URL and xref linting only inspects lines introduced by the PR even when main has advanced since the feature branch diverged.

Changes:

• Switch scripts/lint_urls.sh and scripts/lint_xrefs.sh from git diff base..head to git diff base...head.
• Add a regression test that constructs diverged main and feature histories and asserts only feature-side changes are linted.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

File	Description
scripts/lint_xrefs.sh	Use merge-base diff semantics for PR-only xref linting.
scripts/lint_urls.sh	Use merge-base diff semantics for PR-only URL linting.
.ci/scripts/tests/test_link_check_diff_selection.py	Regression test covering diverged-history selection for both URL and xref linting.

Suppressed comments (2)

scripts/lint_urls.sh:78

• With set -euo pipefail, this per-file pipeline can terminate the whole diff scan when a changed file has no “+” lines (e.g., deletion-only diffs). In that case grep -Ev '^\+\+\+' returns exit code 1, which aborts the process substitution early and can cause the script to miss URLs in later files.

 git diff --unified=0 "$1...$2" -- "$filename" "${excludes[@]}" \
| grep -E '^\+' \
| grep -Ev '^\+\+\+' \
| perl -nle 'print for m#'"$pattern"'#g' \
| sed 's|^|'"$filename"':|'

scripts/lint_xrefs.sh:43

• With set -euo pipefail, this per-file pipeline can terminate the whole diff scan when a changed file has no “+” lines (e.g., deletion-only diffs). In that case grep -Ev '^\+\+\+' returns exit code 1, which aborts the process substitution early and can cause the script to miss xrefs in later files.

 git diff --unified=0 "$1...$2" -- "$filename" "${excludes[@]}" \
| grep -E '^\+' \
| grep -Ev '^\+\+\+' \
| perl -nle 'print for m#'"$pattern"'#g' \
| sed 's|^|'"$filename"':|'

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated no new comments.

Suppressed comments (5)

.github/workflows/_link_check.yml:33

• This workflow still uses the default shallow checkout (actions/checkout without fetch-depth: 0). If lint_urls.sh / lint_xrefs.sh use merge-base diffing (git diff A...B), a shallow clone can be missing the merge base even after git fetch origin "$BASE_REF", causing merge-base-based diffs to fail.

Consider unshallowing the checkout (or setting fetch-depth: 0) so merge-base computations are reliable.

 if [ -n "$BASE_REF" ] && [ "$BASE_REF" != "0000000000000000000000000000000000000000" ]; then
git fetch --no-tags origin "$BASE_REF"
args=("$BASE_REF" "$HEAD_REF")

.ci/scripts/tests/test_link_check_diff_selection.py:104

• Similarly, this xref test won't distinguish BASE..HEAD from BASE...HEAD because main.md is a base-side addition (which becomes a deletion in the diff and is filtered out). Make main remove an xref that exists in the merge base so BASE..HEAD would incorrectly lint it as an added line.

 def test_lint_xrefs_uses_merge_base_for_changed_lines(self) -> None:
base_sha, head_sha = self.create_diverged_history(
"feature.md",
textwrap.dedent(
"""\
[feature](docs/feature-target.md)
"""
),
"main.md",
textwrap.dedent(
"""\
[main](docs/main-target.md)
"""
),
)

.ci/scripts/tests/test_link_check_diff_selection.py:80

• As written, this test won't fail if lint_urls.sh still uses git diff BASE..HEAD: the main.md URL is introduced only on the base side, so it shows up as a deletion (and the script filters to + lines). To make the test distinguish .. vs ..., have main remove a URL that exists in the merge base (so BASE..HEAD would treat it as a new + line in HEAD).

 def test_lint_urls_uses_merge_base_for_changed_lines(self) -> None:
base_sha, head_sha = self.create_diverged_history(
"feature.md",
"https://example.com/feature\n",
"main.md",
"https://example.com/main\n",
)

scripts/lint_xrefs.sh:40

• git diff A...B depends on being able to compute the merge base. In the link-check workflow this script is typically run in an actions/checkout shallow clone (default fetch-depth: 1), and _link_check.yml still fetches BASE_REF shallowly for some jobs, which can leave the merge base unreachable and make this loop fail with "no merge base" errors.

Consider computing the merge base once with git merge-base (so failures can be reported with a clear message) and diffing $merge_base..$2 instead of using ... directly.

 if [ $# -eq 2 ]; then
for filename in $(git diff --name-only --unified=0 "$1...$2"); do
git diff --unified=0 "$1...$2" -- "$filename" "${excludes[@]}" \
| grep -E '^\+' \

.ci/scripts/tests/test_link_check_diff_selection.py:24

• The test setup doesn't currently create any URL/xref in the merge base commit, so it can't detect the difference between git diff BASE..HEAD and git diff BASE...HEAD. To exercise the regression, the merge base should contain a URL/xref that main later removes (so BASE..HEAD would see it as an added line in HEAD, even though it predates the branch).

Without that, these tests can pass even if the scripts still use .. diffing.

This issue also appears in the following locations of the same file:

• line 74
• line 90

 self.write("README.md", "base\n")
self.write("docs/feature-target.md", "feature target\n")
self.write("docs/main-target.md", "main target\n")
self.run_cmd("git add README.md docs/feature-target.md docs/main-target.md")

[psiddh]

Closing this PR in favir of #21765

[Copilot]

Summary

link-check / lint-urls was diffing base..head, so PR jobs could lint unrelated URLs introduced on main after the branch diverged. This narrows both URL and xref linting to lines introduced by the PR by diffing from the merge base instead.

• Scope

  • Update scripts/lint_urls.sh to use git diff base...head
  • Update scripts/lint_xrefs.sh to use the same merge-base semantics

• Behavior change

  • Before: lint jobs could inspect files changed only on the base branch
  • After: lint jobs inspect only files/lines reachable from the PR head relative to the merge base

• Regression coverage

  • Add a focused test that creates diverged main and feature histories and verifies only the feature-side changes are linted for:
    • URL checks
    • cross-reference checks

# before
git diff "$BASE..$HEAD"# after
git diff "$BASE...$HEAD"

Test plan

Added regression coverage in .ci/scripts/tests/test_link_check_diff_selection.py for both lint_urls.sh and lint_xrefs.sh.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/21762

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❌ 2 New Failures

As of commit ed0c8cd with merge base e60faa2 ():

NEW FAILURES - The following jobs have failed:

• Build Presets / apple (pybind) / build (gh)
  ##[error]fatal: unable to access 'https://github.com/leetal/ios-cmake/': SSL certificate problem: self signed certificate
• Lint / link-check / lint-urls (gh)

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[github-actions]

This PR needs a release notes: label

If your change should be included in the release notes (i.e. would users of this library care about this change?), please use a label starting with release notes:. This helps us keep track and include your important work in the next release notes.

To add a label, you can comment to pytorchbot, for example
@pytorchbot label "release notes: none"

For more information, see
https://github.com/pytorch/pytorch/wiki/PyTorch-AutoLabel-Bot#why-categorize-for-release-notes-and-how-does-it-work.

[Copilot]

Pull request overview

This PR updates the link-check lint scripts to diff against the merge base (base...head) instead of the base tip (base..head), so URL and xref linting only inspects lines introduced by the PR even when main has advanced since the feature branch diverged.

Changes:

• Switch scripts/lint_urls.sh and scripts/lint_xrefs.sh from git diff base..head to git diff base...head.
• Add a regression test that constructs diverged main and feature histories and asserts only feature-side changes are linted.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

File	Description
scripts/lint_xrefs.sh	Use merge-base diff semantics for PR-only xref linting.
scripts/lint_urls.sh	Use merge-base diff semantics for PR-only URL linting.
.ci/scripts/tests/test_link_check_diff_selection.py	Regression test covering diverged-history selection for both URL and xref linting.

Suppressed comments (2)

scripts/lint_urls.sh:78

• With set -euo pipefail, this per-file pipeline can terminate the whole diff scan when a changed file has no “+” lines (e.g., deletion-only diffs). In that case grep -Ev '^\+\+\+' returns exit code 1, which aborts the process substitution early and can cause the script to miss URLs in later files.

 git diff --unified=0 "$1...$2" -- "$filename" "${excludes[@]}" \
| grep -E '^\+' \
| grep -Ev '^\+\+\+' \
| perl -nle 'print for m#'"$pattern"'#g' \
| sed 's|^|'"$filename"':|'

scripts/lint_xrefs.sh:43

• With set -euo pipefail, this per-file pipeline can terminate the whole diff scan when a changed file has no “+” lines (e.g., deletion-only diffs). In that case grep -Ev '^\+\+\+' returns exit code 1, which aborts the process substitution early and can cause the script to miss xrefs in later files.

 git diff --unified=0 "$1...$2" -- "$filename" "${excludes[@]}" \
| grep -E '^\+' \
| grep -Ev '^\+\+\+' \
| perl -nle 'print for m#'"$pattern"'#g' \
| sed 's|^|'"$filename"':|'

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated no new comments.

Suppressed comments (5)

.github/workflows/_link_check.yml:33

• This workflow still uses the default shallow checkout (actions/checkout without fetch-depth: 0). If lint_urls.sh / lint_xrefs.sh use merge-base diffing (git diff A...B), a shallow clone can be missing the merge base even after git fetch origin "$BASE_REF", causing merge-base-based diffs to fail.

Consider unshallowing the checkout (or setting fetch-depth: 0) so merge-base computations are reliable.

 if [ -n "$BASE_REF" ] && [ "$BASE_REF" != "0000000000000000000000000000000000000000" ]; then
git fetch --no-tags origin "$BASE_REF"
args=("$BASE_REF" "$HEAD_REF")

.ci/scripts/tests/test_link_check_diff_selection.py:104

• Similarly, this xref test won't distinguish BASE..HEAD from BASE...HEAD because main.md is a base-side addition (which becomes a deletion in the diff and is filtered out). Make main remove an xref that exists in the merge base so BASE..HEAD would incorrectly lint it as an added line.

 def test_lint_xrefs_uses_merge_base_for_changed_lines(self) -> None:
base_sha, head_sha = self.create_diverged_history(
"feature.md",
textwrap.dedent(
"""\
[feature](docs/feature-target.md)
"""
),
"main.md",
textwrap.dedent(
"""\
[main](docs/main-target.md)
"""
),
)

.ci/scripts/tests/test_link_check_diff_selection.py:80

• As written, this test won't fail if lint_urls.sh still uses git diff BASE..HEAD: the main.md URL is introduced only on the base side, so it shows up as a deletion (and the script filters to + lines). To make the test distinguish .. vs ..., have main remove a URL that exists in the merge base (so BASE..HEAD would treat it as a new + line in HEAD).

 def test_lint_urls_uses_merge_base_for_changed_lines(self) -> None:
base_sha, head_sha = self.create_diverged_history(
"feature.md",
"https://example.com/feature\n",
"main.md",
"https://example.com/main\n",
)

scripts/lint_xrefs.sh:40

• git diff A...B depends on being able to compute the merge base. In the link-check workflow this script is typically run in an actions/checkout shallow clone (default fetch-depth: 1), and _link_check.yml still fetches BASE_REF shallowly for some jobs, which can leave the merge base unreachable and make this loop fail with "no merge base" errors.

Consider computing the merge base once with git merge-base (so failures can be reported with a clear message) and diffing $merge_base..$2 instead of using ... directly.

 if [ $# -eq 2 ]; then
for filename in $(git diff --name-only --unified=0 "$1...$2"); do
git diff --unified=0 "$1...$2" -- "$filename" "${excludes[@]}" \
| grep -E '^\+' \

.ci/scripts/tests/test_link_check_diff_selection.py:24

• The test setup doesn't currently create any URL/xref in the merge base commit, so it can't detect the difference between git diff BASE..HEAD and git diff BASE...HEAD. To exercise the regression, the merge base should contain a URL/xref that main later removes (so BASE..HEAD would see it as an added line in HEAD, even though it predates the branch).

Without that, these tests can pass even if the scripts still use .. diffing.

This issue also appears in the following locations of the same file:

• line 74
• line 90

 self.write("README.md", "base\n")
self.write("docs/feature-target.md", "feature target\n")
self.write("docs/main-target.md", "main target\n")
self.run_cmd("git add README.md docs/feature-target.md docs/main-target.md")

[psiddh]

Closing this PR in favir of #21765

[Copilot]

Summary

link-check / lint-urls was diffing base..head, so PR jobs could lint unrelated URLs introduced on main after the branch diverged. This narrows both URL and xref linting to lines introduced by the PR by diffing from the merge base instead.

• Scope

  • Update scripts/lint_urls.sh to use git diff base...head
  • Update scripts/lint_xrefs.sh to use the same merge-base semantics

• Behavior change

  • Before: lint jobs could inspect files changed only on the base branch
  • After: lint jobs inspect only files/lines reachable from the PR head relative to the merge base

• Regression coverage

  • Add a focused test that creates diverged main and feature histories and verifies only the feature-side changes are linted for:
    • URL checks
    • cross-reference checks

# before
git diff "$BASE..$HEAD"# after
git diff "$BASE...$HEAD"

Test plan

Added regression coverage in .ci/scripts/tests/test_link_check_diff_selection.py for both lint_urls.sh and lint_xrefs.sh.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/21762

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❌ 2 New Failures

As of commit ed0c8cd with merge base e60faa2 ():

NEW FAILURES - The following jobs have failed:

• Build Presets / apple (pybind) / build (gh)
  ##[error]fatal: unable to access 'https://github.com/leetal/ios-cmake/': SSL certificate problem: self signed certificate
• Lint / link-check / lint-urls (gh)

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[github-actions]

This PR needs a release notes: label

If your change should be included in the release notes (i.e. would users of this library care about this change?), please use a label starting with release notes:. This helps us keep track and include your important work in the next release notes.

To add a label, you can comment to pytorchbot, for example
@pytorchbot label "release notes: none"

For more information, see
https://github.com/pytorch/pytorch/wiki/PyTorch-AutoLabel-Bot#why-categorize-for-release-notes-and-how-does-it-work.

[Copilot]

Pull request overview

This PR updates the link-check lint scripts to diff against the merge base (base...head) instead of the base tip (base..head), so URL and xref linting only inspects lines introduced by the PR even when main has advanced since the feature branch diverged.

Changes:

• Switch scripts/lint_urls.sh and scripts/lint_xrefs.sh from git diff base..head to git diff base...head.
• Add a regression test that constructs diverged main and feature histories and asserts only feature-side changes are linted.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

File	Description
scripts/lint_xrefs.sh	Use merge-base diff semantics for PR-only xref linting.
scripts/lint_urls.sh	Use merge-base diff semantics for PR-only URL linting.
.ci/scripts/tests/test_link_check_diff_selection.py	Regression test covering diverged-history selection for both URL and xref linting.

Suppressed comments (2)

scripts/lint_urls.sh:78

• With set -euo pipefail, this per-file pipeline can terminate the whole diff scan when a changed file has no “+” lines (e.g., deletion-only diffs). In that case grep -Ev '^\+\+\+' returns exit code 1, which aborts the process substitution early and can cause the script to miss URLs in later files.

 git diff --unified=0 "$1...$2" -- "$filename" "${excludes[@]}" \
| grep -E '^\+' \
| grep -Ev '^\+\+\+' \
| perl -nle 'print for m#'"$pattern"'#g' \
| sed 's|^|'"$filename"':|'

scripts/lint_xrefs.sh:43

• With set -euo pipefail, this per-file pipeline can terminate the whole diff scan when a changed file has no “+” lines (e.g., deletion-only diffs). In that case grep -Ev '^\+\+\+' returns exit code 1, which aborts the process substitution early and can cause the script to miss xrefs in later files.

 git diff --unified=0 "$1...$2" -- "$filename" "${excludes[@]}" \
| grep -E '^\+' \
| grep -Ev '^\+\+\+' \
| perl -nle 'print for m#'"$pattern"'#g' \
| sed 's|^|'"$filename"':|'

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated no new comments.

Suppressed comments (5)

.github/workflows/_link_check.yml:33

• This workflow still uses the default shallow checkout (actions/checkout without fetch-depth: 0). If lint_urls.sh / lint_xrefs.sh use merge-base diffing (git diff A...B), a shallow clone can be missing the merge base even after git fetch origin "$BASE_REF", causing merge-base-based diffs to fail.

Consider unshallowing the checkout (or setting fetch-depth: 0) so merge-base computations are reliable.

 if [ -n "$BASE_REF" ] && [ "$BASE_REF" != "0000000000000000000000000000000000000000" ]; then
git fetch --no-tags origin "$BASE_REF"
args=("$BASE_REF" "$HEAD_REF")

.ci/scripts/tests/test_link_check_diff_selection.py:104

• Similarly, this xref test won't distinguish BASE..HEAD from BASE...HEAD because main.md is a base-side addition (which becomes a deletion in the diff and is filtered out). Make main remove an xref that exists in the merge base so BASE..HEAD would incorrectly lint it as an added line.

 def test_lint_xrefs_uses_merge_base_for_changed_lines(self) -> None:
base_sha, head_sha = self.create_diverged_history(
"feature.md",
textwrap.dedent(
"""\
[feature](docs/feature-target.md)
"""
),
"main.md",
textwrap.dedent(
"""\
[main](docs/main-target.md)
"""
),
)

.ci/scripts/tests/test_link_check_diff_selection.py:80

• As written, this test won't fail if lint_urls.sh still uses git diff BASE..HEAD: the main.md URL is introduced only on the base side, so it shows up as a deletion (and the script filters to + lines). To make the test distinguish .. vs ..., have main remove a URL that exists in the merge base (so BASE..HEAD would treat it as a new + line in HEAD).

 def test_lint_urls_uses_merge_base_for_changed_lines(self) -> None:
base_sha, head_sha = self.create_diverged_history(
"feature.md",
"https://example.com/feature\n",
"main.md",
"https://example.com/main\n",
)

scripts/lint_xrefs.sh:40

• git diff A...B depends on being able to compute the merge base. In the link-check workflow this script is typically run in an actions/checkout shallow clone (default fetch-depth: 1), and _link_check.yml still fetches BASE_REF shallowly for some jobs, which can leave the merge base unreachable and make this loop fail with "no merge base" errors.

Consider computing the merge base once with git merge-base (so failures can be reported with a clear message) and diffing $merge_base..$2 instead of using ... directly.

 if [ $# -eq 2 ]; then
for filename in $(git diff --name-only --unified=0 "$1...$2"); do
git diff --unified=0 "$1...$2" -- "$filename" "${excludes[@]}" \
| grep -E '^\+' \

.ci/scripts/tests/test_link_check_diff_selection.py:24

• The test setup doesn't currently create any URL/xref in the merge base commit, so it can't detect the difference between git diff BASE..HEAD and git diff BASE...HEAD. To exercise the regression, the merge base should contain a URL/xref that main later removes (so BASE..HEAD would see it as an added line in HEAD, even though it predates the branch).

Without that, these tests can pass even if the scripts still use .. diffing.

This issue also appears in the following locations of the same file:

• line 74
• line 90

 self.write("README.md", "base\n")
self.write("docs/feature-target.md", "feature target\n")
self.write("docs/main-target.md", "main target\n")
self.run_cmd("git add README.md docs/feature-target.md docs/main-target.md")

[psiddh]

Closing this PR in favir of #21765

[Copilot]

Summary

link-check / lint-urls was diffing base..head, so PR jobs could lint unrelated URLs introduced on main after the branch diverged. This narrows both URL and xref linting to lines introduced by the PR by diffing from the merge base instead.

• Scope

  • Update scripts/lint_urls.sh to use git diff base...head
  • Update scripts/lint_xrefs.sh to use the same merge-base semantics

• Behavior change

  • Before: lint jobs could inspect files changed only on the base branch
  • After: lint jobs inspect only files/lines reachable from the PR head relative to the merge base

• Regression coverage

  • Add a focused test that creates diverged main and feature histories and verifies only the feature-side changes are linted for:
    • URL checks
    • cross-reference checks

# before
git diff "$BASE..$HEAD"# after
git diff "$BASE...$HEAD"

Test plan

Added regression coverage in .ci/scripts/tests/test_link_check_diff_selection.py for both lint_urls.sh and lint_xrefs.sh.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/21762

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❌ 2 New Failures

As of commit ed0c8cd with merge base e60faa2 ():

NEW FAILURES - The following jobs have failed:

• Build Presets / apple (pybind) / build (gh)
  ##[error]fatal: unable to access 'https://github.com/leetal/ios-cmake/': SSL certificate problem: self signed certificate
• Lint / link-check / lint-urls (gh)

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[github-actions]

This PR needs a release notes: label

If your change should be included in the release notes (i.e. would users of this library care about this change?), please use a label starting with release notes:. This helps us keep track and include your important work in the next release notes.

To add a label, you can comment to pytorchbot, for example
@pytorchbot label "release notes: none"

For more information, see
https://github.com/pytorch/pytorch/wiki/PyTorch-AutoLabel-Bot#why-categorize-for-release-notes-and-how-does-it-work.

[Copilot]

Pull request overview

This PR updates the link-check lint scripts to diff against the merge base (base...head) instead of the base tip (base..head), so URL and xref linting only inspects lines introduced by the PR even when main has advanced since the feature branch diverged.

Changes:

• Switch scripts/lint_urls.sh and scripts/lint_xrefs.sh from git diff base..head to git diff base...head.
• Add a regression test that constructs diverged main and feature histories and asserts only feature-side changes are linted.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

File	Description
scripts/lint_xrefs.sh	Use merge-base diff semantics for PR-only xref linting.
scripts/lint_urls.sh	Use merge-base diff semantics for PR-only URL linting.
.ci/scripts/tests/test_link_check_diff_selection.py	Regression test covering diverged-history selection for both URL and xref linting.

Suppressed comments (2)

scripts/lint_urls.sh:78

• With set -euo pipefail, this per-file pipeline can terminate the whole diff scan when a changed file has no “+” lines (e.g., deletion-only diffs). In that case grep -Ev '^\+\+\+' returns exit code 1, which aborts the process substitution early and can cause the script to miss URLs in later files.

 git diff --unified=0 "$1...$2" -- "$filename" "${excludes[@]}" \
| grep -E '^\+' \
| grep -Ev '^\+\+\+' \
| perl -nle 'print for m#'"$pattern"'#g' \
| sed 's|^|'"$filename"':|'

scripts/lint_xrefs.sh:43

• With set -euo pipefail, this per-file pipeline can terminate the whole diff scan when a changed file has no “+” lines (e.g., deletion-only diffs). In that case grep -Ev '^\+\+\+' returns exit code 1, which aborts the process substitution early and can cause the script to miss xrefs in later files.

 git diff --unified=0 "$1...$2" -- "$filename" "${excludes[@]}" \
| grep -E '^\+' \
| grep -Ev '^\+\+\+' \
| perl -nle 'print for m#'"$pattern"'#g' \
| sed 's|^|'"$filename"':|'

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated no new comments.

Suppressed comments (5)

.github/workflows/_link_check.yml:33

• This workflow still uses the default shallow checkout (actions/checkout without fetch-depth: 0). If lint_urls.sh / lint_xrefs.sh use merge-base diffing (git diff A...B), a shallow clone can be missing the merge base even after git fetch origin "$BASE_REF", causing merge-base-based diffs to fail.

Consider unshallowing the checkout (or setting fetch-depth: 0) so merge-base computations are reliable.

 if [ -n "$BASE_REF" ] && [ "$BASE_REF" != "0000000000000000000000000000000000000000" ]; then
git fetch --no-tags origin "$BASE_REF"
args=("$BASE_REF" "$HEAD_REF")

.ci/scripts/tests/test_link_check_diff_selection.py:104

• Similarly, this xref test won't distinguish BASE..HEAD from BASE...HEAD because main.md is a base-side addition (which becomes a deletion in the diff and is filtered out). Make main remove an xref that exists in the merge base so BASE..HEAD would incorrectly lint it as an added line.

 def test_lint_xrefs_uses_merge_base_for_changed_lines(self) -> None:
base_sha, head_sha = self.create_diverged_history(
"feature.md",
textwrap.dedent(
"""\
[feature](docs/feature-target.md)
"""
),
"main.md",
textwrap.dedent(
"""\
[main](docs/main-target.md)
"""
),
)

.ci/scripts/tests/test_link_check_diff_selection.py:80

• As written, this test won't fail if lint_urls.sh still uses git diff BASE..HEAD: the main.md URL is introduced only on the base side, so it shows up as a deletion (and the script filters to + lines). To make the test distinguish .. vs ..., have main remove a URL that exists in the merge base (so BASE..HEAD would treat it as a new + line in HEAD).

 def test_lint_urls_uses_merge_base_for_changed_lines(self) -> None:
base_sha, head_sha = self.create_diverged_history(
"feature.md",
"https://example.com/feature\n",
"main.md",
"https://example.com/main\n",
)

scripts/lint_xrefs.sh:40

• git diff A...B depends on being able to compute the merge base. In the link-check workflow this script is typically run in an actions/checkout shallow clone (default fetch-depth: 1), and _link_check.yml still fetches BASE_REF shallowly for some jobs, which can leave the merge base unreachable and make this loop fail with "no merge base" errors.

Consider computing the merge base once with git merge-base (so failures can be reported with a clear message) and diffing $merge_base..$2 instead of using ... directly.

 if [ $# -eq 2 ]; then
for filename in $(git diff --name-only --unified=0 "$1...$2"); do
git diff --unified=0 "$1...$2" -- "$filename" "${excludes[@]}" \
| grep -E '^\+' \

.ci/scripts/tests/test_link_check_diff_selection.py:24

• The test setup doesn't currently create any URL/xref in the merge base commit, so it can't detect the difference between git diff BASE..HEAD and git diff BASE...HEAD. To exercise the regression, the merge base should contain a URL/xref that main later removes (so BASE..HEAD would see it as an added line in HEAD, even though it predates the branch).

Without that, these tests can pass even if the scripts still use .. diffing.

This issue also appears in the following locations of the same file:

• line 74
• line 90

 self.write("README.md", "base\n")
self.write("docs/feature-target.md", "feature target\n")
self.write("docs/main-target.md", "main target\n")
self.run_cmd("git add README.md docs/feature-target.md docs/main-target.md")

[psiddh]

Closing this PR in favir of #21765

[Copilot]

Summary

link-check / lint-urls was diffing base..head, so PR jobs could lint unrelated URLs introduced on main after the branch diverged. This narrows both URL and xref linting to lines introduced by the PR by diffing from the merge base instead.

• Scope

  • Update scripts/lint_urls.sh to use git diff base...head
  • Update scripts/lint_xrefs.sh to use the same merge-base semantics

• Behavior change

  • Before: lint jobs could inspect files changed only on the base branch
  • After: lint jobs inspect only files/lines reachable from the PR head relative to the merge base

• Regression coverage

  • Add a focused test that creates diverged main and feature histories and verifies only the feature-side changes are linted for:
    • URL checks
    • cross-reference checks

# before
git diff "$BASE..$HEAD"# after
git diff "$BASE...$HEAD"

Test plan

Added regression coverage in .ci/scripts/tests/test_link_check_diff_selection.py for both lint_urls.sh and lint_xrefs.sh.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/21762

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❌ 2 New Failures

As of commit ed0c8cd with merge base e60faa2 ():

NEW FAILURES - The following jobs have failed:

• Build Presets / apple (pybind) / build (gh)
  ##[error]fatal: unable to access 'https://github.com/leetal/ios-cmake/': SSL certificate problem: self signed certificate
• Lint / link-check / lint-urls (gh)

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[github-actions]

This PR needs a release notes: label

If your change should be included in the release notes (i.e. would users of this library care about this change?), please use a label starting with release notes:. This helps us keep track and include your important work in the next release notes.

To add a label, you can comment to pytorchbot, for example
@pytorchbot label "release notes: none"

For more information, see
https://github.com/pytorch/pytorch/wiki/PyTorch-AutoLabel-Bot#why-categorize-for-release-notes-and-how-does-it-work.

[Copilot]

Pull request overview

This PR updates the link-check lint scripts to diff against the merge base (base...head) instead of the base tip (base..head), so URL and xref linting only inspects lines introduced by the PR even when main has advanced since the feature branch diverged.

Changes:

• Switch scripts/lint_urls.sh and scripts/lint_xrefs.sh from git diff base..head to git diff base...head.
• Add a regression test that constructs diverged main and feature histories and asserts only feature-side changes are linted.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

File	Description
scripts/lint_xrefs.sh	Use merge-base diff semantics for PR-only xref linting.
scripts/lint_urls.sh	Use merge-base diff semantics for PR-only URL linting.
.ci/scripts/tests/test_link_check_diff_selection.py	Regression test covering diverged-history selection for both URL and xref linting.

Suppressed comments (2)

scripts/lint_urls.sh:78

• With set -euo pipefail, this per-file pipeline can terminate the whole diff scan when a changed file has no “+” lines (e.g., deletion-only diffs). In that case grep -Ev '^\+\+\+' returns exit code 1, which aborts the process substitution early and can cause the script to miss URLs in later files.

 git diff --unified=0 "$1...$2" -- "$filename" "${excludes[@]}" \
| grep -E '^\+' \
| grep -Ev '^\+\+\+' \
| perl -nle 'print for m#'"$pattern"'#g' \
| sed 's|^|'"$filename"':|'

scripts/lint_xrefs.sh:43

• With set -euo pipefail, this per-file pipeline can terminate the whole diff scan when a changed file has no “+” lines (e.g., deletion-only diffs). In that case grep -Ev '^\+\+\+' returns exit code 1, which aborts the process substitution early and can cause the script to miss xrefs in later files.

 git diff --unified=0 "$1...$2" -- "$filename" "${excludes[@]}" \
| grep -E '^\+' \
| grep -Ev '^\+\+\+' \
| perl -nle 'print for m#'"$pattern"'#g' \
| sed 's|^|'"$filename"':|'

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated no new comments.

Suppressed comments (5)

.github/workflows/_link_check.yml:33

• This workflow still uses the default shallow checkout (actions/checkout without fetch-depth: 0). If lint_urls.sh / lint_xrefs.sh use merge-base diffing (git diff A...B), a shallow clone can be missing the merge base even after git fetch origin "$BASE_REF", causing merge-base-based diffs to fail.

Consider unshallowing the checkout (or setting fetch-depth: 0) so merge-base computations are reliable.

 if [ -n "$BASE_REF" ] && [ "$BASE_REF" != "0000000000000000000000000000000000000000" ]; then
git fetch --no-tags origin "$BASE_REF"
args=("$BASE_REF" "$HEAD_REF")

.ci/scripts/tests/test_link_check_diff_selection.py:104

• Similarly, this xref test won't distinguish BASE..HEAD from BASE...HEAD because main.md is a base-side addition (which becomes a deletion in the diff and is filtered out). Make main remove an xref that exists in the merge base so BASE..HEAD would incorrectly lint it as an added line.

 def test_lint_xrefs_uses_merge_base_for_changed_lines(self) -> None:
base_sha, head_sha = self.create_diverged_history(
"feature.md",
textwrap.dedent(
"""\
[feature](docs/feature-target.md)
"""
),
"main.md",
textwrap.dedent(
"""\
[main](docs/main-target.md)
"""
),
)

.ci/scripts/tests/test_link_check_diff_selection.py:80

• As written, this test won't fail if lint_urls.sh still uses git diff BASE..HEAD: the main.md URL is introduced only on the base side, so it shows up as a deletion (and the script filters to + lines). To make the test distinguish .. vs ..., have main remove a URL that exists in the merge base (so BASE..HEAD would treat it as a new + line in HEAD).

 def test_lint_urls_uses_merge_base_for_changed_lines(self) -> None:
base_sha, head_sha = self.create_diverged_history(
"feature.md",
"https://example.com/feature\n",
"main.md",
"https://example.com/main\n",
)

scripts/lint_xrefs.sh:40

• git diff A...B depends on being able to compute the merge base. In the link-check workflow this script is typically run in an actions/checkout shallow clone (default fetch-depth: 1), and _link_check.yml still fetches BASE_REF shallowly for some jobs, which can leave the merge base unreachable and make this loop fail with "no merge base" errors.

Consider computing the merge base once with git merge-base (so failures can be reported with a clear message) and diffing $merge_base..$2 instead of using ... directly.

 if [ $# -eq 2 ]; then
for filename in $(git diff --name-only --unified=0 "$1...$2"); do
git diff --unified=0 "$1...$2" -- "$filename" "${excludes[@]}" \
| grep -E '^\+' \

.ci/scripts/tests/test_link_check_diff_selection.py:24

• The test setup doesn't currently create any URL/xref in the merge base commit, so it can't detect the difference between git diff BASE..HEAD and git diff BASE...HEAD. To exercise the regression, the merge base should contain a URL/xref that main later removes (so BASE..HEAD would see it as an added line in HEAD, even though it predates the branch).

Without that, these tests can pass even if the scripts still use .. diffing.

This issue also appears in the following locations of the same file:

• line 74
• line 90

 self.write("README.md", "base\n")
self.write("docs/feature-target.md", "feature target\n")
self.write("docs/main-target.md", "main target\n")
self.run_cmd("git add README.md docs/feature-target.md docs/main-target.md")

[psiddh]

Closing this PR in favir of #21765

[Copilot]

Summary

link-check / lint-urls was diffing base..head, so PR jobs could lint unrelated URLs introduced on main after the branch diverged. This narrows both URL and xref linting to lines introduced by the PR by diffing from the merge base instead.

• Scope

  • Update scripts/lint_urls.sh to use git diff base...head
  • Update scripts/lint_xrefs.sh to use the same merge-base semantics

• Behavior change

  • Before: lint jobs could inspect files changed only on the base branch
  • After: lint jobs inspect only files/lines reachable from the PR head relative to the merge base

• Regression coverage

  • Add a focused test that creates diverged main and feature histories and verifies only the feature-side changes are linted for:
    • URL checks
    • cross-reference checks

# before
git diff "$BASE..$HEAD"# after
git diff "$BASE...$HEAD"

Test plan

Added regression coverage in .ci/scripts/tests/test_link_check_diff_selection.py for both lint_urls.sh and lint_xrefs.sh.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/21762

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❌ 2 New Failures

As of commit ed0c8cd with merge base e60faa2 ():

NEW FAILURES - The following jobs have failed:

• Build Presets / apple (pybind) / build (gh)
  ##[error]fatal: unable to access 'https://github.com/leetal/ios-cmake/': SSL certificate problem: self signed certificate
• Lint / link-check / lint-urls (gh)

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[github-actions]

This PR needs a release notes: label

If your change should be included in the release notes (i.e. would users of this library care about this change?), please use a label starting with release notes:. This helps us keep track and include your important work in the next release notes.

To add a label, you can comment to pytorchbot, for example
@pytorchbot label "release notes: none"

For more information, see
https://github.com/pytorch/pytorch/wiki/PyTorch-AutoLabel-Bot#why-categorize-for-release-notes-and-how-does-it-work.

[Copilot]

Pull request overview

This PR updates the link-check lint scripts to diff against the merge base (base...head) instead of the base tip (base..head), so URL and xref linting only inspects lines introduced by the PR even when main has advanced since the feature branch diverged.

Changes:

• Switch scripts/lint_urls.sh and scripts/lint_xrefs.sh from git diff base..head to git diff base...head.
• Add a regression test that constructs diverged main and feature histories and asserts only feature-side changes are linted.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

File	Description
scripts/lint_xrefs.sh	Use merge-base diff semantics for PR-only xref linting.
scripts/lint_urls.sh	Use merge-base diff semantics for PR-only URL linting.
.ci/scripts/tests/test_link_check_diff_selection.py	Regression test covering diverged-history selection for both URL and xref linting.

Suppressed comments (2)

scripts/lint_urls.sh:78

• With set -euo pipefail, this per-file pipeline can terminate the whole diff scan when a changed file has no “+” lines (e.g., deletion-only diffs). In that case grep -Ev '^\+\+\+' returns exit code 1, which aborts the process substitution early and can cause the script to miss URLs in later files.

 git diff --unified=0 "$1...$2" -- "$filename" "${excludes[@]}" \
| grep -E '^\+' \
| grep -Ev '^\+\+\+' \
| perl -nle 'print for m#'"$pattern"'#g' \
| sed 's|^|'"$filename"':|'

scripts/lint_xrefs.sh:43

• With set -euo pipefail, this per-file pipeline can terminate the whole diff scan when a changed file has no “+” lines (e.g., deletion-only diffs). In that case grep -Ev '^\+\+\+' returns exit code 1, which aborts the process substitution early and can cause the script to miss xrefs in later files.

 git diff --unified=0 "$1...$2" -- "$filename" "${excludes[@]}" \
| grep -E '^\+' \
| grep -Ev '^\+\+\+' \
| perl -nle 'print for m#'"$pattern"'#g' \
| sed 's|^|'"$filename"':|'

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated no new comments.

Suppressed comments (5)

.github/workflows/_link_check.yml:33

• This workflow still uses the default shallow checkout (actions/checkout without fetch-depth: 0). If lint_urls.sh / lint_xrefs.sh use merge-base diffing (git diff A...B), a shallow clone can be missing the merge base even after git fetch origin "$BASE_REF", causing merge-base-based diffs to fail.

Consider unshallowing the checkout (or setting fetch-depth: 0) so merge-base computations are reliable.

 if [ -n "$BASE_REF" ] && [ "$BASE_REF" != "0000000000000000000000000000000000000000" ]; then
git fetch --no-tags origin "$BASE_REF"
args=("$BASE_REF" "$HEAD_REF")

.ci/scripts/tests/test_link_check_diff_selection.py:104

• Similarly, this xref test won't distinguish BASE..HEAD from BASE...HEAD because main.md is a base-side addition (which becomes a deletion in the diff and is filtered out). Make main remove an xref that exists in the merge base so BASE..HEAD would incorrectly lint it as an added line.

 def test_lint_xrefs_uses_merge_base_for_changed_lines(self) -> None:
base_sha, head_sha = self.create_diverged_history(
"feature.md",
textwrap.dedent(
"""\
[feature](docs/feature-target.md)
"""
),
"main.md",
textwrap.dedent(
"""\
[main](docs/main-target.md)
"""
),
)

.ci/scripts/tests/test_link_check_diff_selection.py:80

• As written, this test won't fail if lint_urls.sh still uses git diff BASE..HEAD: the main.md URL is introduced only on the base side, so it shows up as a deletion (and the script filters to + lines). To make the test distinguish .. vs ..., have main remove a URL that exists in the merge base (so BASE..HEAD would treat it as a new + line in HEAD).

 def test_lint_urls_uses_merge_base_for_changed_lines(self) -> None:
base_sha, head_sha = self.create_diverged_history(
"feature.md",
"https://example.com/feature\n",
"main.md",
"https://example.com/main\n",
)

scripts/lint_xrefs.sh:40

• git diff A...B depends on being able to compute the merge base. In the link-check workflow this script is typically run in an actions/checkout shallow clone (default fetch-depth: 1), and _link_check.yml still fetches BASE_REF shallowly for some jobs, which can leave the merge base unreachable and make this loop fail with "no merge base" errors.

Consider computing the merge base once with git merge-base (so failures can be reported with a clear message) and diffing $merge_base..$2 instead of using ... directly.

 if [ $# -eq 2 ]; then
for filename in $(git diff --name-only --unified=0 "$1...$2"); do
git diff --unified=0 "$1...$2" -- "$filename" "${excludes[@]}" \
| grep -E '^\+' \

.ci/scripts/tests/test_link_check_diff_selection.py:24

• The test setup doesn't currently create any URL/xref in the merge base commit, so it can't detect the difference between git diff BASE..HEAD and git diff BASE...HEAD. To exercise the regression, the merge base should contain a URL/xref that main later removes (so BASE..HEAD would see it as an added line in HEAD, even though it predates the branch).

Without that, these tests can pass even if the scripts still use .. diffing.

This issue also appears in the following locations of the same file:

• line 74
• line 90

 self.write("README.md", "base\n")
self.write("docs/feature-target.md", "feature target\n")
self.write("docs/main-target.md", "main target\n")
self.run_cmd("git add README.md docs/feature-target.md docs/main-target.md")

[psiddh]

Closing this PR in favir of #21765